SRE and Platform Engineer
Ritesh Sonawane
I build reliable Kubernetes, observability, CI/CD, and multi-cloud systems for production teams.
I have helped 10+ clients across AWS, GCP, and Azure scale infra across 17+ Kubernetes clusters, including GPU-enabled and air-gapped enterprise environments.
Startups I've worked with
Selected posts
Recent writing
Jul 18, 2026 · 7 min read
Types of LLM Explained: Dense, MoE, and Everything In Between
LLM types from dense transformers to MoE, reasoning models, and embedding models, covering what each one actually does differently and when to use it. Written for anyone who wants to understand model architecture beyond just the benchmark name.
Jul 6, 2026 · 11 min read
LLM Inference KPIs Every SRE Should Know
LLM inference metrics for SREs: TTFT, TPOT, KV cache, HBM bandwidth, and how they connect to real production behavior.
Jun 26, 2026 · 10 min read
Kubernetes Gateway API Inference Extension
Kubernetes Gateway API Inference Extension: What It Is and Why It Matters